Posts

Showing posts with the label operating expenses

AI Data Centers: A Fragile Business Model Requiring Perfect Execution & A Lot Of Luck

Image
1. Massive capital expenditures. Much higher capital intensity per rack than a standard data center.  2. Extremely high operating costs due to the power intensity of AI infrastructure. 3. Low customer switching costs shift negotiating power from supplier to customer.  4. Huge gap between capex and revenue flows: 2 to 5 years. 5. Heavy dependence on debt financing and hence vulnerability to rising risk premiums and interest rates. 6. Underlying technology, large language models, is fundamentally flawed according to many computer scientists. This is corroborated by persistent hallucinations and weak problem solving abilities. 7. Large language models show sharply diminishing returns to scale. 8. Luck is required because there are only a few, very large customers for these projects and loss of a single customer can kill a project. Luck is also important because any of these data center projects requires upgrades to the power transmission grid and also generator capacity. 

Oracle's Downfall: A Bad Bet On AI Data Centers

Image
AI data centers has two challenges: big capex requirements and high operating expenses due to the skyrocketing cost of electricity and the power intensive nature of GPUs. Another challenge is low customer switching costs. A customer like OpenAI or Anthropic rarely puts equipment into an AI data center. It simply downloads its software into the data center's bare metal servers, estimates its model, and then uploads the estimated model back to its own servers. In Oracle's case, it has only one major customer and hence faces huge customer risk. AI data center contracts tend to be soft and often include no penalty cancellation clauses with 60 or 90 days prior notice. This AI data center business model only works if the vendor has access to cheap electricity like hydropower in the Nordics and lots of  customers as oppose to one or two large ones. Solar and wind power do not generally make the grade because their fully loaded pricing is in fact quite high. In the case of Oracle, the ...